Abstract Klebsiella pneumoniae is a leading cause of hospital-acquired infections in the United States and the most common Carbapenem-resistant Enterobacteriaceae (CRE) and Extended-Spectrum Beta-lactamase (ESBL) species. Infections with CRE cause up to 50% mortality from sepsis, and both ESBL and CRE infections are a significant cause of excess morbidity and hospital costs. Our preliminary data from 1765 patients indicates that patients with K. pneumoniae gastrointestinal colonization are at a high risk of subsequent disease (Odds ratio 4.0; p<0.0001) and become infected with their colonizing strain. Antibiotic therapy can be life-saving but choosing the correct regimen requires antimicrobial susceptibility data that is available days after the onset of disease. Testing for colonization could provide an ideal opportunity for intervention: physicians can identify at- risk patients and use antibiotic susceptibility data from their colonizing strain to make rational choices for empiric therapy. High-risk patients could be targeted for intervention, but how the complex interaction of patient and bacterial factors leads to disease is unknown. To close this gap in knowledge, we have assembled a multi- disciplinary team of physician-scientists, epidemiologists, bioinformaticians, and statisticians with expertise in clinical microbiology, microbial pathogenesis and infectious diseases. The objective of this proposal is to identify the bacterial and host factors that predict K. pneumoniae infections in colonized patients. Our central hypothesis is that K. pneumoniae strains vary in their virulence potential, and the combination of K. pneumoniae genotype and host susceptibility determines the risk of disease in a colonized patient. To test this hypothesis, we validated a novel genome comparison method called Pathogenicity-Associated Loci sequencing (PAL-Seq) to identify K. pneumoniae genes in variable genomic regions that are associated with infection. We also developed a preliminary clinical model of patient risk factors for K. pneumoniae infection. We will test our hypothesis and meet the objective of this proposal through the following specific aims: Aim 1: Define patient risk factors for K. pneumoniae infection in colonized patients. We will use electronic medical records and culture samples in cohorts from three hospitals to build and validate models based on patient characteristics and colonization density as risk factors for infection in colonized patients, and test the models in the subgroup of ESBL and CRE colonized patients. Aim 2: Identify K. pneumoniae genes that predict the risk of disease in colonized patients. Using colonizing and invasive isolates, we will apply our PAL-Seq pipeline to identify bacterial genes associated with infection, validate them in animal models and an independent cohort, and test candidate virulence genes in ESBL and CRE colonized patients. The positive impact of this work will be immediate and substantial. We will rapidly advance our understanding of K. pneumoniae pathogenesis based on both clinical and animal studies, and develop predictive models that could be used to identify high- risk patients for prevention or rapid treatment of K. pneumoniae infection.